The performance of artificial intelligence in the exams of tourist guidance

Abstract

The aim of this study is to evaluate the efficiency of ChatGPT versions 3.5 and 4 for training tourist guides. The study followed a systematic approach by conducting assessments on undergraduate students from three institutions who are enrolled in tourist guide education programs and both ChatGPT versions. Competent academicians assessed a succession of questions in the form of open-ended and multiple-choice questions. The mean scores obtained on the multiple-choice test for ChatGPT-4 were better than those of both ChatGPT-3.5 and college students, thereby indicating that ChatGPT-4 has greatly improved. Nevertheless, when responding to open-ended queries, individuals with real-life experience as tour guides gave much more inclusive as well as convincing answers compared to ChatGPT-4. This underscores the importance of hands-on experiences in training tour guides, where AI technology is currently weak. This study contributes to better comprehension regarding the role played by artificial intelligence (AI) in education with reference to the tourism industry specifically. While at the same time emphasizing how critical human expertise is needed during practical learning sessions, this implies that AI has potential for disseminating theoretical knowledge. The results suggest that AI is a beneficial supplementary aid in educational environments, rather than a replacement for human-centered instructional approaches.